This respository also includes high-resolution versions of the paper figures; figures 2-12 were generated with ggplot2 (see plot_paper_figures.R) which were generated with ggplot2.
Maintained by Matthew Brehmer.
-
Figure 1: Linear and Radial temperature range charts designed for mobile phone displays, representative of the stimuli used in our crowdsourced experiment. The colored bars encode observed temperature ranges and are superimposed on gray bars encoding average temperature ranges. Corresponding Week, Month, and Year charts display the same data (Seattle temperatures in 2015).
-
Table 1: Our experiment included five experimental tasks, presented to participants in this order, from relatively simple to complex. This table illustrates these tasks using Sleep range data. Each task was preceded by an introductory instruction (top row). The second row shows an example trial for each task. The table also indicates the possible response values and the number of trials for each task. To complement this table, we provide a video depicting the experiment as supplemental material.
-
Figure 2: Mean completion Times in seconds for tasks 1-5 with Radial and Linear layouts. Task abbreviations: T1-LD = Locate Date, T2-RV = Read Value, T3-LM = Locate Min / Max, T4-CV = Compare Values, T5-CR = Compare Ranges. Orange = Temperature; Blue = Sleep. As indicated in Table 1, T2 and T4 did not include trials at the Year level. Error bars are 95% CIs; for guidance with respect to interpreting overlaps in CIs, refer to M. Krzywinski and N. Altman. Points of significance: Error bars. Nature Methods, 10:921–922, 2013
-
Figure 3: Error Rates for tasks 1-5 with Radial and Linear layouts.
-
Figure 4: Proportion of participants who prefer either a Radial or Linear layout at each granularity (Week / Month / Year).
-
Figure 5: Participants’ reported confidence from 1 (low) to 5 (high) with Radial and Linear layouts at each granularity (Week / Month / Year).
-
Figure 6: Radial / Linear Completion Time ratios by task. In cases where a CI intersects the dashed line (a ratio of 1), we interpret this as inconclusive evidence for an effect.
-
Figure 7: Radial / Linear Completion Time ratios by granularity.
-
Figure 8: Radial − Linear Error Rate differences by task. In cases where a CI intersects the dashed line (a difference of 0%), we interpret this as inconclusive evidence for an effect.
-
Figure 9: Pairwise Year, Month, and Week Completion Time ratios.
-
Figure 10: Pairwise Year, Month, and Week Error Rate differences.
-
Figure 11: Radial / Linear Completion Time ratios by target range value.
-
Figure 12: Radial − Linear Error Rate differences by target range value.